Use of Multiple Regression Model to Relate Waste Composition with Energy Content of Refuse-Derived Fuel
Chee, Thiam Ming (2006) Use of Multiple Regression Model to Relate Waste Composition with Energy Content of Refuse-Derived Fuel. Masters thesis, Universiti Putra Malaysia.
The amount of wastes in Malaysia has increased tremendously as a result of rapid economic growth. Currently, landfill is the only method for the disposal of solid wastes in the country. It was estimated that about 16,000 tons of municipal solid wastes ended up in landfill everyday. This scenario has worsen when existing landfills are fast filling up and the possibility of getting a new landfill is becoming more difficult because of land scarcity and increase of land prices especially in the urban areas. Further, as the environmental impact of existing landfills has started to become more apparent, the government is under pressure to find a better solution to dispose the wastes without having to incur excessive costs. This study was thus initiated as part of the effort to examine the feasibility of converting solid wastes into refusederived- fuel (RDF). The objectives of this study were to determine the relationship between the different waste composition and the calorific value of Refuse Derived Fuel in order to develop models to estimate the calorific value (energy output) of Rehse Derived Fuel based on the specific waste composition. The results from this study revealed that food waste was the biggest portion in Kajang (44%) followed by plastic (13.2%), paper (10.5%) and yard waste (9.3%). The correlation indicated that food wastes, yard wastes and paper wastes were correlated negatively with the calorific value while plastic is the only waste component having positive correlations with RDF calorific value. Three non-linear regression models were developed from the experimental results. The first model is found to be LCVwet(kcal/kg) = 10 (3.4643 + 0.001736 Fd + 0.005781 PIS - 0.007268 M) , where Fd is food waste, Pls is plastics, and M is moisture. The adjusted R2 for this model was 0.8706. The second model is found as LCVwet (kcalkg) = 10 (0.57013 + 0.02618 M + 0.03375 VM) where M is moisture, and VM is volatile matter. The adjusted R~ for this model was 0.6163. The final model developed from this study is LCV* (kcam) = (0.044549 - 0.000283 C + 0.000168 0 ~ ~ )an' d~ th'e ~ind ependent variable was carbon (C) and oxygen (0) content. The adjusted R2 was 0.8227. From this study, it can be concluded that calorific value can be estimated through regression model based on physical composition, ultimate and proximate analysis, and the models were also more accurate than models developed by other researchers for solid waste.
Repository Staff Only: Edit item detail